Master's Research Work

Flame Acceleration in an Obstacle-Laden Tube

M. Tech Research Project

Advisor - Prof. Amit Kumar and Prof. Satyanarayanan Chakravarthy

Brief Description

Studied premixed flame acceleration in an obstructed channel to understand deflagration-to-detonation transition (DDT) phenomenon in pulsed detonation engines (PDEs).
Designed and constructed an experimental apparatus (square cross-sectional tube with closely spaced obstacles) to validate the laminar flame acceleration mechanism proposed by Vitaly Bychkov of Umea University.
Adopted Schlieren, Shadowgraphy, and High-Speed Imaging technique to visualize the flame front and Image Processing techniques to compute flame propagation speed and acceleration.
Interpreted the results and observed the flame front not accelerating monotonically throughout the obstructed channel which contradicted the analytical results.
Concluded that the reduction of flame tip speed seen at the later stages of the channel can possibly occur because of the heat loss to the obstacles and tube walls.

Flame Acceleration by Shchelkin Mechanism

Thermal expansion of combustion products behind the flame front induces a flow in the unburned gas. (Fig. 1)
Non-uniform flow field generated by no-slip condition at the tube walls increases the flame surface area. (Fig. 1)
The interaction of flame front with large scale structures in unburned gas past the obstacles leads to global flame area enhancement. (known as flame folding) (Fig. 2)
Small scale turbulence generated in the shear layer and the re-circulation zone enhances the local burning velocity. (Fig. 3)
Compression waves generated by flame coalesce to form shock waves which interact with flame front causing severe flame distortion. (Fig. 4)

Fig. 1. A schematic diagram showing gas expansion. (Image courtesy - Ref [4])

Fig. 2. Flame folding mechanism. (Image courtesy - Ref [1])

Fig. 3. Re-circulation zone and shear layer formation in the unburned gas flow. (Image courtesy - Ref [1])

Fig. 4. Shock-flame interaction. (Image courtesy - Ref [2])

Laminar Flame Acceleration Mechanism

General belief

Flame acceleration is caused solely due to the turbulence generated by the obstacles in the unburned gas flow.
Acceleration is more pronounced when the inter-obstacle spacing is of the order of channel height.

Work by Bychkov et al. (2008) [3]

Stated that turbulence was a fatal obstacle for quantitative understanding of flame acceleration phenomenon.
Proposed a physical mechanism of flame acceleration which can quantify the acceleration.
Gas expansion due to the delayed burning of reaction mixture in the pockets between the obstacles produces a strong jet flow in the unobstructed part of the tube. (Fig. 5)
The jet flow renders the flame tip to propagate much faster which produces new pockets, generates a positive feedback between the flame and the flow.
This phenomenon can be observed in channels with closely spaced obstacles where flame acceleration contribution from turbulence is minimal.

Fig. 5. (a) Obstacle characteristics and flame front boundary in channels with closely spaced obstacles. (b) Production of jet flow in the central unobstructed part due to the expansion of product gases in the vertical channels. (Image courtesy Ref [4])

Fig. 6. Theoretical formulation for laminar flame acceleration mechanism in obstructed channels proposed by Bychkov et al. (2008). The flame tip location is an exponential function in time and the flame tip velocity is a linear function in distance. (Ref [3] and [4])

Flame Acceleration Apparatus (Combustion Tube)

config_1.pdf

Fig. 7. Obstacle characteristics in an obstructed channel.

config_new_1.pdf

Fig. 8. CAD model of the combustion tube assembly.

config_new_3.pdf

Fig. 9. 3-D schematic of the window module created using Bender software.

config_new_2.pdf

Fig. 10. Picture of the window module.

config_new_4.pdf

Fig. 11. 3-D schematic of the window module. (Isometric view)

config_3.pdf

Fig. 12. 3-D schematic of the window module. The cut section view shows the obstructed part of the channel with multiple orifice plates.

config_2.pdf

Fig. 13. Orifice plate obstacles.

config_new_5.pdf

Fig. 14. The orifice plate assembly.

config_new_6.pdf

Fig. 15. Combustion tube assembly mounted on a stand.

Experimental Setup

setup_new_2.pdf

Fig. 16. Photograph of ignition transformer.

setup_new_3.pdf

Fig. 17. Photograph of vacuum pump.

setup_new_4.pdf

Fig. 18. Photograph showing vacuum line (brown hose), capillary tubes for supplying reaction mixture to the combustion tube, electronic pressure gauge, and igniter.

setup_new_1.pdf

Fig. 19. Gas flow circuit.

setup_new_11.pdf

Fig. 20. Control panel showing the capillary connections between combustion tube and the mixing chamber.

setup_new_5.pdf

Fig. 21. Schematic diagram of Schlieren setup.

setup_new_6.pdf

Fig. 22. Light source and slit arrangement for Schlieren imaging.

setup_new_7.pdf

Fig. 23. Knife edge (a razor blade).

setup_new_9.pdf

Fig. 24. A 470 mm focal length Convex lens.

setup_new_12.pdf

Fig. 25. Photron high-speed star CMOS camera to capture Schlieren images.

setup_new_10.pdf

Fig. 26. A 30 cm diameter and 3 m focal length parabolic mirror.

setup_new_8.pdf

Fig. 27. A 100 watt halogen lamp.

setup_new_13.pdf

Fig. 28. Alignment of camera, convex lens, knife edge, and parabolic mirror for Schlieren experiment.

setup_new_16.pdf

Fig. 29. Teflon casing for mounting pressure transducers.

setup_new_17.pdf

Fig. 30. Data acquisition system (DAQ) and Signal conditioner.

setup_new_18.pdf

Fig. 31. Igniter synchronization circuit.

config_4.pdf

Fig. 32. The full experimental apparatus: combustion tube, control panel, air connection, methane cylinder etc. Electronic pressure gauges, non-return valves are also shown.

Experimental Results

Fig. 33, 34, and 35 shows flame acceleration in the initial 50 cm of the obstructed channel.
The field of view in Fig. 33 starts from 6 cm and ends at 32 cm from the ignition end with a inter-frame time of 1 ms.
Time taken to traverse the entire field of view in Fig. 33 is 7 ms.
In this case, the window module shown in Fig. 10 which has a total length of 50 cm is placed adjacent to the igniter.

Res_1_1.pdf

Fig. 33. Schlieren video showing development of flame surface.

Res_1_2.pdf

Fig. 34. Time history of flame front location obtained by taking average for six independent experiments. Error bars are also plotted.

Res_1_3.pdf

Fig. 35. Time history of flame tip speed with 'actual curve' i.e. flame tip speed from one image frame to the next and 'fitted curve' which is the time derivative of a 4th order polynomial fitted curve for flame location data.

Fig. 36, 37, and 38 show the flame tip speed as a function of distance from the ignition end with experimental data and a 4th order polynomial fitted curve.
Here, the window module shown in Fig. 10 which has a total length of 50 cm is shifted to 3 different locations to obtain the flame speed data throughout the entire obstructed length (0 to 2 m).
In Fig. 33, window module is kept at first position i.e. the module starts from 0 cm and ends at 50 cm from the ignition end.

Res_2_1.pdf

Fig. 36. Flame speed data with window module in 2nd position. (50 cm to 100 cm)

Res_2_2.pdf

Fig. 37. Flame speed data with window module in 3rd position. (100 cm to 150 cm)

Res_2_3.pdf

Fig. 38. Flame speed data with window module in 4th position. (150 cm to 200 cm)

Fig. 39, 40, 41, and 42 compare the experimental results for flame tip speed vs distance with analytical formulation of laminar flame acceleration proposed Bychkov et al. (2008) [3].
It is observed that the flame tip speed decreases (Except when window module is in 1st position) in the initial part of the window module irrespective of its position.
A gap between the obstacle plates and the glass wall is considered to be the possible reason. This problem is only present in the window module. In all the other modules (Metal modules), the obstacle plates are housed inside the metal grooves at its edges.
So, the flame tip speeds recorded at the beginning of the window in all 4 positions is used to obtain a predicted velocity profile (through curve fitting) in an obstructed channel which doesn’t have gaps between the wall and the obstacles or any other altered boundary conditions. The image is shown in Fig. 43.
It is observed from the predicted velocity profile that the flame front is not accelerating monotonically throughout the obstructed channel which contradict the analytical results.
The reduction in flame tip speed seen at the later stages of the channel can possibly occur because of the heat loss to the obstacles and channel walls.

Res_3_1.pdf

Fig. 39. Flame speed data with window module in 1st position. (0 cm to 50 cm)

Res_3_2.pdf

Fig. 40. Flame speed data with window module in 2nd position. (50 cm to 100 cm)

Res_3_3.pdf

Fig. 41. Flame speed data with window module in 3rd position. (100 cm to 150 cm)

Res_3_4.pdf

Fig. 42. Flame speed data with window module in 4th position. (150 cm to 200 cm)

Res_3_5.pdf

Fig. 43. Flame tip speed vs distance plot with experimental data obtained throughout the entire obstructed channel (0 to 2m) and the predicted velocity profile.

Flame Pictures

Flame_Pic_1.pdf

Fig. 44. Images of flame in channels with closely spaced orifice plates. Pictures are captured through a 30 FPS mobile camera.

Flame_Pic_2.pdf

Fig. 45. Images of flame in channels with widely spaced orifice plates. Pictures are captured through a 30 FPS mobile camera.

Flame Acceleration Videos

Video 1. Flame propagation in the initial 50 cm from the ignition end of the flame acceleration apparatus. Schlieren flow visualization technique is used to track the flame front.

Video 2. Zoomed version of the video 1 showing formation of vortices in the fuel-air mixture pockets in between the obstacles.

References

[1] Craig T. Johansen and Gaby Ciccarelli, "Visualization of the unburned gas flow field ahead of an accelerating flame in an obstructed square channel ", Combustion and Flame, 156(2) (2009), 405–416.

[2] Gaby Ciccarelli, Craig T. Johansen, and Michael Parravani, "The role of shock–flame interactions on flame acceleration in an obstacle laden channel", Combustion and Flame, 157(2010), 2125–2136.

[3] Vitaly Bychkov, Damir Valiev, and Lars-Erik Eriksson,“Physical mechanism of ultrafast flame acceleration”, Phys.Rev.Lett.101, 164501 (2008).

[4] Damir Valiev, Vitaly Bychkov, V’yacheslav Akkerman, Chung K. Law, and Lars-Erik Eriksson, "Flame acceleration in channels with obstacles in the deflagration-to-detonation transition", Combustion and Flame, 157 (2010), 1012–1021.

If you wish to know more about the results or any technical details, please contact me personally. You can do that by clicking here.

Google Sites

Report abuse